Science Friday
Hosted by Flora Lichtman
A bioethicist walks us through the complexity of clinical trials, from their core scientific questions to patient selection to FDA approval.
3,496 words · 4 speakers
Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
Hey, it's Flora and you are listening to Science Friday. We're kicking this conversation off with a story.
Science is personal, and when you have a chronic illness like I do, the drugs that you get make a huge difference.
This is Chris. He's from Florida, and he called us because he tried to get into a clinical trial, and his experience raised some questions. Chris has autoimmune arthritis. Over the years, he's taken 10 different drugs, and they all worked, but each one for less time than the one before.
And when the drugs work, you really do feel like those commercials. You're dancing around and leaping off of the playground into the swimming pool. And then when they stop working, you get really sad.
So Chris tried to enroll in a clinical trial for a new drug for his condition.
Because new drugs are being developed all the time. And I thought, oh, this will be perfect for me.
But Chris was rejected from participating, and he was rejected because he was on his 10th drug, which of course makes him exactly the kind of person who might benefit most from a new drug.
And I really wondered what you're supposed to do if you're one of the problem children who doesn't meet the eligibility criteria, because on one level, who needs a clinical trial more than the guy who's, quote unquote, failing his 10th drug, or having the 10th drug fail him, depending on how you want to phrase it?
So that is what we're looking into today: clinical trials and how they work. Are there incentives for drug developers to leave out, quote, "problem children"? Or is it more complicated than that? Here with us is Dr. Holly Fernandez Lynch, a lawyer and a bioethicist who studies the ethics of clinical trials and the FDA's pharmaceutical policies at the University of Pennsylvania. Holly, welcome to Science Friday.
Thanks so much for having me.
Let's get to Chris's experience. I mean, is this the norm? That drug developers exclude people from trials who have, quote, "failed other drugs"?
So yes, it's very typical that when you're doing a clinical trial, you need to specify what the inclusion criteria and the exclusion criteria are. That's part of doing good science because what you're trying to figure out is whether the drug works and you want to be able to attribute the outcomes that you're seeing to that intervention rather than to extraneous factors that might influence the results. So inclusion criteria are of course going to be things like you have the disorder of interest and exclusion criteria could be a wide variety of things, including your age, maybe your BMI, your ability to speak English and consent on your own behalf.
And it would also typically include things like comorbidities, other diseases that you have that could influence the study outcomes. And related to that might be other drugs that you have previously tried. So if you're like the gentleman in our example here, if you've tried seven or 10 different products, those might actually have changed the course of your disease in a particular way that would make your result different from somebody who's trying a drug for the very first time.
Interesting. So it's about getting clean data.
Exactly, right? So what we're trying to do when we're doing a clinical trial is to get out all of the extraneous detail to really drill down to see, is this result coming because of the intervention? On the other hand, we have to take a step back and ask, what is the most important research question? That question might be, how can we get a drug that's going to work for patients for whom other drugs have failed? That's a different scientific question than, does this drug work? And so we might think in the context of oncology, you see this a lot, where there's different lines of therapy. Your first line therapy fails and then you go to your second or third or even fourth line therapies.
So when new oncology drugs are being studied, they're typically being studied for later line therapies, right? After others have failed. And then, once they get approval for that indication, you can keep studying them so that they can kind of move up the list. Eventually, the ideal is that they would be the first drug that a patient tried. But you would design your study completely different depending on what scientific question you're asking.
That makes a ton of sense. I mean, this seems so basic, but does a clinical trial have to prove a drug is effective to get greenlit by the FDA?
Yeah, so this is a really important question to understand how drugs get regulated in the US. But imagine you have a drug that's never been approved for any use before. What you need to demonstrate to FDA is that it's safe for its intended use, that it's effective for its intended use, and that you can manufacture it appropriately and consistently so you're getting the same drug every single time. So that's a pretty basic question, right? Does this drug work for people who have this disease? It's not required for FDA approval that you show that your drug is better than another drug that's already been approved or is available to patients.
In other words, if there's five other drugs on the market, you could design a clinical trial that asks, does this drug outcompete? Does it do better than these other existing drugs? But you don't have to ask that. You could just ask, does it do anything regardless of how it compares to the other existing drugs?
That's right, although unsurprisingly, it's a little bit of an oversimplification. And I'll just give you an example, right? Imagine that there are several other drugs that are already on the market, and it would be unsafe to ask the people who are in your study to stop taking those existing drugs in order to take your new drug. You're not going to be able to get that study off the ground for ethical reasons. But from a regulatory perspective, really, it's just a very straightforward question. Does this drug work better than nothing, right? That's the typical design.
Does it work better than nothing? Right.
And so you might see, right, a study that's evaluating a product against a placebo. But more often, right, what you're seeing is a drug being studied against what's called the standard of care. So what would a doctor typically provide a patient who has this disorder? Right, you're going to typically be testing your new drugs against that baseline for very pragmatic reasons. You're not going to be able to get people to enroll in the study and you're not going to be able to get your ethics approval if you don't have an appropriate standard of care. But that's more, that turns out to be more complicated in practice. There can be debates about what's the appropriate standard of care.
Hmm, I mean, and, and I think there is some nuance about what it means to be effective, right? Like, aren't there cases in which drugs are approved that seem to help, but they don't maybe improve the patient's experience of the condition?
There's so much in that question that you, that you just asked. There's lots of different ways that you can meet that standard of effectiveness. The best way is to show that this drug has a positive impact on how the patient feels, functions, or survives. Those are things that we typically refer to as clinical endpoints. You don't need a laboratory test, you don't need a scan, right? The patient will be able to tell, I'm feeling better. And over, you know, the course of looking at many patients, you would see a different outcome in their, in their disease. That is in contrast to what is referred to as surrogate endpoints.
Surrogate endpoints, they, they don't directly measure how the patient feels, functions, or survives. Instead, they measure outcomes that are reasonably likely to predict the thing that we actually care about. So I'll give you an example from, from the HIV context. We know that if we can influence a patient's viral load or their CD4 counts, that is going to have an impact on their disease. Course, those are called validated surrogates. We know exactly how they relate to the outcomes of interest. There are lots of other surrogates that have not been validated, but we think they're likely to predict benefit.
So for example, you might in the oncology space, see progression free survival, right? How long do you survive without your disease progressing? Or you might see things like tumor shrinkage. Or in the Alzheimer's case, you might have seen in the news, some discussion about plaques in the brain, right? These are things that you would need to have a lab test or a scan to see whether they're being affected by the, by the drug. But we think that they're predictive. We're just not a hundred percent sure. So FDA can grant approval of drugs based on those unvalidated surrogate measures. I know that's a lot of jargon, but basically FDA can grant approval
based on the prediction of benefit. And when it does that, it requires companies to continue studying the product after it's allowed on the market. So, so that's one type of flexibility that FDA has, where you can get a drug on the market where we're not even really sure that it works, but we hope it does. It has a lot of discretion in what evidence it's willing to accept to meet that effectiveness standard.
I mean, this Alzheimer's drug that you mentioned, there was some controversy about it getting approved, right? For this reason.
Just a tad. Yeah, yeah. Okay, so, so the Alzheimer's drug that you're referring to is a drug called Adjuhelm. It's actually been voluntarily taken off the market by, by its manufacturer, Biogen. But it was quite controversial. It was approved in 2021 using this accelerated approval pathway that I was just describing. So it was approved not based on a demonstrated benefit in Alzheimer's disease outcomes, but rather based on the influence of the drug on these plaques in the brain. It was controversial because both that endpoint seemed to be a bit dicey, right?
There, there was a lot of discussion about whether it was even reasonably likely to predict clinical benefit. The company had also tested that drug against the clinical endpoints, those things about how the patient feels, functions, or survives. And it turned out that the drug wasn't able to demonstrate benefit on any of those. The company relied on these plaques in the brain as kind of a backup to get the product approved. FDA is able to call on outside advisors, outside experts, to weigh in on these kind of challenging drug approval decisions. And they called this advisory committee together. Everybody but one person who abstained said to FDA, no, the standard has not been met
here. Do not approve this drug. And FDA went ahead and approved it anyway, despite a lot of concern about whether or not it worked. So it's really kind of held up as, I was going to say a shining example, but whatever the opposite of shining is, the tarnished example of the kind of drug that really should not be approved. But there are other circumstances where FDA grants flexible approvals, you know, where it's the evidence is a bit uncertain. But when you're facing a serious disease with no good treatment options, it makes sense for FDA to be able to exert a bit more flexibility. We just want to make sure that that flexibility doesn't go overboard so that we can have
confidence that when our doctor prescribes us an FDA-approved drug, that it is reasonably likely to benefit us based on strong evidence.
We got to take a break, but when we come back, we're going to talk about who oversees the design of clinical trials. Stay with us.
Support for Science Friday comes from Science Sandbox, an initiative of the Simons Foundation.
Okay, Holly, who oversees the design of clinical trials?
So there are a lot of different parties involved here, right? So you might hear folks talk about a drug manufacturer or a drug sponsor, right? That's the company that is trying to get its product through the FDA approval pathway. So that sponsor, before they can even start testing their drug in human beings, they have to go to FDA and file an application and say, here's what we're planning to do, here's how we're going to protect the research participants, here's our study design and why we think it's going to generate the evidence that you need FDA to decide whether we've met the standard for approval. And FDA will grant, you know, grant approval of that application, allowing the company to begin doing their tests in people.
The other party that's involved is what's called an institutional review board or an IRB. But really what their responsibility is, is to look after the rights and welfare of the people who are going to be enrolling in a research study. So the IRB is evaluating the study based on the risks and benefits. It wants to make sure that the benefits are reasonably likely to outweigh the risks. The benefits that we're talking about, though, are not necessarily limited to benefits to the research participants. Because the reason we do research is actually to generate knowledge for the benefit of other people, right? Future patients who are not in that study.
So some IRBs are run by universities, others are independent or commercially run, and some of which are even backed by private equity. Do these different types of IRBs change how they approach the review process?
The main difference between an academic IRB and one of these commercial IRBs is their, their structure. So at a university IRB, there will be a committee of volunteers, right? So it'll be some professors from that university and some outside individuals who are, who are included on the committee. But the key difference is that the commercial IRBs... It is the it is the sole job of those IRB members to review protocols. Right. They're not also, you know, responsible for teaching and research and other service to their university. This is all they do, right? It's their full time. It's their full time work.
So there's a benefit to that, right? You kind of build up a lot of expertise. You have more resources, more policies. You know, you can improve your ability to review research in that way. The downside, right, is if you think about who is the customer of any IRB? It is not the research participant. The research participant has no say whatsoever in which IRB is going to be reviewing their the study that they participate in. But keep in mind, the IRB's whole job is to protect the participants. Who is picking the IRB, though, is going to be the sponsor.
Does that mean like, are there IRBs that are known to sort of move faster than other IRBs? Is there sort of differences in the way that the ethics are interpreted by different IRBs?
Yeah, so here's here's what I can tell you. There have not been robust comparisons of different types of IRBs and the outcomes that they achieve for a wide variety of reasons, right? It becomes a bit complicated. You need to understand a bit about the regulations. They are pretty bare bones, right? I was describing before. Risks and benefits are reasonable. Okay, what does that mean? Two different IRBs could interpret that in much different ways. They would both meet the regulatory standard, but one might kind of go above and beyond and make sure that there's even stronger ethical protections in place. And the other might say, look, this is all the regulations require. You meet it. You're compliant. Go on your go on your merry way.
So if what you're selling is rapid turnaround time and regulatory compliance, it's more likely that you would fall on the minimum side of the spectrum, right? And it's more likely that somebody who had more freedom and flexibility with their investigators would say, well, we know the regulations don't require you to do this, but we think Judgment does require you to go a bit above and beyond. That's the kind of difference that we're talking about here. It's it's quite nuanced and very, very difficult to kind of study IRBs side by side, which is a problem. These these entities should be much more transparent, given the power that they have to decide what research proceeds.
You know, so Marty McCary is out at FDA. We have seen this FDA sort of take up the mantle for some treatments, cast doubt on other medical interventions. Has anything changed about how drugs are approved under this administration?
Well, the statutory standard remains the same, right? I feel like a bit of a broken record, but the statutory standard is so simple, right? Is the drug safe and effective for its intended use? The sorts of things that have changed at FDA in the past year or so have to do with a couple of things. One is, how quickly is FDA engaging in its review? A major change under, you know, this this administration has been a program called the the Commissioner's Priority Review Vouchers, in which FDA could grant to companies a voucher in which they promised to review the application in only two months, rather than six. And you might think, you know, four months, that's not that big of a deal.
It's a huge deal to companies, right? Four months longer on the market of a blockbuster drug is a a lot of money. And four months fewer to review an application for FDA staff is really, really challenging. And it's important to recognize how FDA engages in its review. They don't just take the company's word for it. They get all of the data that's supporting that product, all the clinical trial data. The FDA is running its own, you know, analyses of those data, really checking that every T is crossed, every I is dotted, so that we can trust the products that are allowed on the market. So the shorter time that the FDA staffers have to review, the more worried we all should be.
Be that maybe something important is going to be missed either around safety or effectiveness. And the other thing I'll add that's changed is FDA is a challenging agency, right? If you think about its role, it's a public health agency. It needs to be science driven, right? But it's not exclusively science driven in the sense that just because a drug has a positive outcome in its in its trial, that doesn't necessarily mean that FDA should grant approval, right? There are judgment calls about approving new drugs because very rarely are you getting such amazing trial results that everyone agrees this is, you know, a stupendous amazing drug and we absolutely should approve it.
So when you have those judgment calls, that can lead to the possibility of political interference. In previous administrations, it's really just been up to the FDA staff reviewers to decide whether a product should get approval without the political appointees kind of dipping their hands in those approval decisions. And that was for a very good reason. Nobody wants to be offered a drug in the doctor's office that they have to question. Well, does this drug actually work and did FDA approve it because the data were good? Or did FDA approve it because the sponsors were buddies with the FDA commissioner or with the president, right?
It's it's really important to maintain strong kind of separation between those approval decisions and any political interference. And we've seen a lot more political interference with this administration than we've seen in the past.
Holly, thanks for being here today.
Thanks again.
Dr. Holly Fernandez Lynch is an associate professor of medical ethics and health policy at the University of Pennsylvania. This episode was produced by Shashana Buxbaum. Thank you very much to listener Chris. Thanks, Chris, for sharing your story and sparking this conversation. And Chris wanted to thank Sabrina, who's helped him navigate the healthcare system. And thank you all for listening. And if you have a question you want us to look into, 8774-SCIFER is our number. We listen to every voicemail and we love hearing from you. I'm Flora Lichtman, and we'll catch you next time.
Page updated October 30, 2026 · transcript and summary are machine-generated